There are two ways to read a pitch deck. A founder reads it forward, as a story: here is a large problem, here is our answer, here is the proof it is working, here is the money we need to finish. An engineer reads it the way they read a design doc from a team they do not yet trust, slide by slide, treating each one as a claim that has to survive contact with the evidence sitting next to it. We have been on both sides. We wrote a deck for a raster-log software company and pitched it, and we have sat across from other people's decks with the quiet, slightly rude habit of a reviewer who wants to see the arithmetic. This note applies that second habit to our own slides, because the most honest specimen to dissect is one where we already know where the bodies are.
The procedure is small and the same on every slide. Decode the slide into three parts. First, the claim: the single thing the slide wants the room to believe. Second, the evidence: the numbers and logos actually printed on it, no more. Third, the hidden assumption: the thing the claim rests on that the slide does not say out loud, because saying it would invite exactly the question an engineer is about to ask. The move that makes this more than skepticism is the last step. You reconstruct the unit-economics bridge the slide left implicit, the division that connects the headline to a countable thing, and you say the result out loud. The slide either survives its own arithmetic or it quietly does not, and the derived number is where you find out.
The market slide claims a share and hides a headcount
Take the market slide first, because it is the one built to impress and the one an engineer discounts fastest. Ours claimed a 134 billion dollar total market, a 6.7 billion dollar serviceable slice, and a 3 percent obtainable share by the end of year five, which resolved to 180 million dollars of revenue. The evidence printed on the slide is that chain of three numbers and the multiplication that links the last two. The claim is that the prize is enormous and a sliver of it is a real company.
The hidden assumption is buried in the word share. A percentage of a market sounds like something you take, as if the 3 percent were a wedge sliced off a pie. It is not. It is a headcount you sell to, one contract at a time, and the slide never prints the headcount. So an engineer prints it. At 1,200 dollars per seat per year, 180 million dollars of revenue is roughly 150,000 paying seats. That changes the question from is the market big enough, which is rhetorical, to can this team put a product in front of 150,000 paying users in five years, which is answerable and much harder. The customer-development literature has said for a decade that a market claim is a hypothesis about who will buy, not an asset on the balance sheet [1]. The engineer's version of that warning is arithmetic: divide the revenue by the price and the abstraction turns into a count you either believe or you do not.
The pricing slide claims a line and hides a curve
The pricing slide is where the market claim goes to be checked, and usually is not, because the two slides are read minutes apart and never held against each other. Ours claimed 1,200 dollars per user per year and, elsewhere, the same 180 million dollars by year five. The evidence is those two numbers. The claim is that the price point is set and the model scales to the target.
The hidden assumption is that seats accumulate in a straight line and, once sold, stay sold. Reconstruct the bridge and the line becomes a curve you have to defend. If year five is 150,000 seats and the first year is 6 million dollars, which is 5,000 seats at the same price, then the plan is a thirtyfold climb in paying seats over four years, with no churn eating the base underneath the growth. That is what the pricing slide hides behind its clean per-seat figure. The lean-analytics case is that a company reduces to a handful of driver metrics, and that headline totals are exactly where the per-unit numbers that govern the model go to hide [2]. A thirtyfold seat climb is not impossible, but it is a specific, aggressive claim about sales velocity and retention that the slide made without appearing to make any claim at all. The engineer's job is to unhide it, then ask what has to be true about the sales motion for the curve to bend that hard.
The traction slide claims proof and hides a conversion
The traction slide is the emotional peak of most decks and where the gap between claim and evidence is widest. Ours showed 3 Letters of Intent, a 6 million dollar first-year projection, and a 3 million dollar seed carrying 18 months of runway. The evidence is three non-binding letters and three numbers. The claim is that demand is proven and the first year is within reach.
Two bridges live under this slide, and an engineer builds both. The first connects the letters to the revenue. Three Letters of Intent are not three customers; they are three parties who have said, in writing that binds no one, that they are interested. The first-year revenue needs roughly 5,000 paid seats, and the honest question is what conversion rate turns 3 expressions of interest into 5,000 paid seats, over what sales cycle. The slide does not say, because the letters are doing rhetorical work as proof when they are really evidence of a pipeline's first inch. The second bridge connects the seed to the clock. A 3 million dollar raise across 18 months of runway is an implied burn of about 167,000 dollars a month, the metronome the whole traction story has to keep time with. Venture-deal mechanics are explicit that a round buys a fixed runway, and the runway sets the pace at which everything else has to happen [3]. So the traction slide, decoded, is a race: 3 letters have to become thousands of seats before an 18-month clock runs out. It is still a fine slide. It is just no longer a proof; it is a bet with a stopwatch, which is what it always was.
Why the derived number is the whole technique
This reading works because a slide is built to make you look at the evidence and feel the claim, and never quite compute the bridge between them out loud. The bridge is where the assumption lives, and the assumption is where the risk is. The three bridges we built, roughly 150,000 seats behind the market target, a thirtyfold seat climb behind the pricing line, and a 167,000 dollar monthly burn behind the traction proof, are not gotchas. They are the deck's own claims, restated in units a person can argue with. We were the founder, so we mean this fairly: reconstructing the bridges does not prove the deck wrong, it proves the deck is a set of claims rather than facts, which is what every deck is and what a good one admits under questioning. A deck that survives having its bridges printed is stronger for it. One that falls apart the moment someone says the seat number out loud was always going to fall apart; the only question was whether it happened in the room or two years later in the bank statement.
The decode, in brief
- Decode every slide into three parts: the claim it wants believed, the evidence actually printed on it, and the hidden assumption the claim rests on. The technique is to reconstruct the unit-economics bridge the slide left implicit and say the resulting number out loud.
- The market slide claims a share and hides a headcount. A 3 percent slice of a 6.7 billion dollar market that yields 180 million dollars at 1,200 dollars a seat is really a claim about selling roughly 150,000 paying seats, which is the number the slide never prints.
- The pricing slide claims a line and hides a curve. If year five is 150,000 seats and year one is 6 million dollars, or about 5,000 seats, the plan is a thirtyfold seat climb over four years with no churn, an aggressive claim disguised as a clean per-seat price.
- The traction slide claims proof and hides a conversion and a clock. Three non-binding LOIs have to become the roughly 5,000 paid seats a 6 million dollar year needs, before a 3 million dollar seed over 18 months of runway, an implied 167,000 dollars a month, runs out.
- The derived number is the whole point. It restates the deck's own claims in units a person can argue with, moving the decision from the impressive total the slide chose to the reachability of the implied count, where the risk actually lives.
Limitations
This is a reading procedure demonstrated on one deck, not a scoring rubric. The sourced figures are real archive numbers: the 134 billion and 6.7 billion dollar markets, the 3 percent share, the 180 million dollar target, the 1,200 dollar seat price, the 6 million dollar first year, the 3 million dollar seed, the 18 months of runway, and the 3 Letters of Intent. The bridge numbers, the roughly 150,000 seats, the thirtyfold climb, the 5,000 first-year seats, and the 167,000 dollar monthly burn, are plain arithmetic on those figures, so they inherit every simplification in them: a single flat price with no tiers or discounts, no expansion revenue, and no churn, which is exactly the assumption an engineer would flag and go ask about. Decoding a slide tells you what question to ask; it does not answer it. Whether 150,000 seats is reachable, whether a 30x climb is credible for this team, and whether 3 letters convert at any useful rate are judgments the arithmetic sets up and cannot settle. The procedure finds the load-bearing assumption faster. It is not a substitute for diligence on whether the assumption holds.
The habit worth keeping
What this reading leaves you with is not cynicism about decks but a reflex about numbers. Whenever a slide hands you a large total, ask what countable thing it divides into, and print that thing. A market becomes a headcount, a revenue target becomes a seat count, a runway becomes a monthly burn, a page of proof becomes a conversion rate with a deadline. None of those translations is hostile. They are just the claim, said in units you can hold. The deck you can trust is the one whose author already did the division before you asked, had the derived number ready, and did not flinch when you said it out loud.
References
[1] Blank, S., and Dorf, B. The Startup Owner's Manual: The Step-by-Step Guide for Building a Great Company. K&S Ranch (2012). The customer-development case that a deck's market and traction claims are hypotheses to be tested rather than facts already banked. https://www.wiley.com/en-us/The+Startup+Owner%27s+Manual%3A+The+Step+By+Step+Guide+for+Building+a+Great+Company-p-9781119690689
[2] Croll, A., and Yoskovitz, B. Lean Analytics: Use Data to Build a Better Startup Faster. O'Reilly (2013). The argument that a business reduces to a few driver metrics, and that headline totals are where the per-unit numbers that govern the model go to hide. https://www.oreilly.com/library/view/lean-analytics/9781449335687/
[3] Feld, B., and Mendelson, J. Venture Deals: Be Smarter Than Your Lawyer and Venture Capitalist, 3rd ed. Wiley (2016). The investor-side account of what a seed round buys and how runway follows from the raise, the frame that turns a seed-and-runway pair into an implied monthly burn. https://www.wiley.com/en-us/Venture+Deals%3A+Be+Smarter+Than+Your+Lawyer+and+Venture+Capitalist%2C+3rd+Edition-p-9781119259756
